This paper asks whether a neat trick for finding the order of neutrino masses stays reliable if there is physics beyond the Standard Model. The trick uses a precision reactor experiment called JUNO and accelerator experiments (T2K and NOvA) that measure how muon neutrinos disappear. By comparing the two kinds of measurements with percent-level precision, researchers had shown this “sum rule” can reveal whether neutrino masses follow the normal or inverted order. The authors check whether new types of interactions could change the measurements enough to flip that conclusion.

The team wrote down how the sum rule is altered if new physics produces a common shift in the effective atmospheric mass-squared value seen by reactor and long-baseline experiments. They used a simple but accurate statistical estimate to combine the long-baseline best-fit values (for normal ordering ∆m2_31 ≈ 2.516×10^-3 eV^2 and for inverted ordering |∆m2_32| ≈ 2.485×10^-3 eV^2, with a long-baseline precision of about 0.031×10^-3 eV^2) and allowed JUNO’s best-fit and precision to vary. JUNO is expected to reach percent or sub-percent precision within months and, when combined with T2K and NOvA, could determine the ordering at roughly the 3-sigma level under standard assumptions. The authors find that a new-physics-induced shift in the mass-squared values across experiments of order a few times 10^-5 eV^2 (they quote a critical scale near 4×10^-5 eV^2) can reverse which ordering the combined data would favor.

To show how this can happen in real models, the paper studies two examples. First, scalar non-standard interactions (SNSI), where a new light scalar alters neutrino masses in matter. Current experimental limits on SNSI couplings make any change to the sum rule too small to matter. Second, neutrinos coupled to an ultralight scalar field can produce time-varying or environment-dependent mass shifts of the size needed to affect the sum rule. In that case, the combined analysis of JUNO and long-baseline data could point to the wrong mass ordering unless the possibility of such a shift is considered.